Identifying affordances from online product reviews

Tianjun Hou , Bernard Yannou (1) , Yann Leroy (1) , Emilie Poirson (2, 3) , Ivan Mata (4) , Georges Fadel (4)

1 LGI - Laboratoire Génie Industriel - EA 2606
2 LS2N - Laboratoire des Sciences du Numérique de Nantes
3 LS2N - équipe PACCE - Perception, Action, Cognition pour la Conception et l’Ergonomie
4 CEDAR

Affordance based design is developed since the beginning of 21 st century. Affordances being revealed properties of a system in a context, they may be much diverse and unexpected. Consequently, it is an utopia to think of enumerating all the existing precise affordances in advance. Presently, identifying affordances along a design or redesign process is based on experiments and focus groups, which are time and resource consuming. Although automatic identification strategies have been proposed, the lack of affordance database along with clear categorization technique makes it unpracticable and non-repeatable today. In this paper, the theoretical basis and technical basis of identifying affordances from online reviews are discussed. A framework of affordance identification is proposed by capturing constitutive affordance elements with natural language processing algorithms. Meanwhile, a case study of 303 review sentences of Kindle Paperwhite from Amazon.com is conducted with 1 expert in affordance based design and 6 participants. The result shows that the proposed framework is effective in affordance identification. It provides basis for automating the identification process in the future.

Domaines: Ingénierie assistée par ordinateur; Autre; Intelligence artificielle [cs.AI]; Modélisation et simulation

Tianjun Hou, Bernard Yannou, Yann Leroy, Emilie Poirson, Ivan Mata, et al.. Identifying affordances from online product reviews. International Conference on Engineering Design (ICED), Aug 2017, Vancouver, Canada. ⟨hal-01673545⟩